SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

The cancer stem cell (CSC) model, in which a small population of cells within a tumor possesses the ability to self-renew and reconstitute the phenotype of primary tumor, has gained wide acceptance based on evidence over the past decade. It has also been reported that cancer cell lines contain a CSC subpopulation. However, phenotypic differences between CSCs and non-CSCs in cancer cell lines are not better defined than in primary tumors. Furthermore, some cell lines do not have a CSC population, revealed as a side population and expression of CD133. Thus, the identification of CSCs in cancer cell lines remains elusive. Here, we investigated the CSC hierarchy within HCT116 colon cancer cells, which do not have a CD133-positive subpopulation. We examined the expression of alternative CSC markers epithelial specific antigen (ESA) and CD44 in floating-sphere-derived cells, which are known to be the cells of enriching CSCs. Sphere-derived HCT116 cells exhibited heterogeneous expression of ESA and CD44. The two major subpopulations of HCT116 sphere cells (ESAlowCD44−/low and ESAhighCD44high) exhibited a biological/proliferative hierarchy of sphere-forming and soft agar colony-forming activity. However, there was no difference between the two subpopulations in the incidence of xenograft tumors. When ESAlowCD44−/low cells were allowed to aggregate and re-form floating-spheres, the biological/proliferative hierarchy of parental HCT116 spheres was reconstituted, in terms of ESA and CD44 expression. Thus, HCT116 cells have plasticity when they are set in floating-spheres, suggesting that maintenance of the HCT116 cell line conforms to a stochastic model, not a CSC model. (Cancer Sci 2009; 100: 2275–2282)

The existence of cancer stem cells (CSCs) in human primary solid tumors, such as glioblastoma, and breast, colon, and hematopoietic tumors, has been confirmed through the use of techniques such as cell fractionation, which is based on the expression of CSC-specific markers, and xenotransplantation.(1–3) The definitive properties of CSCs in primary tumors are broadly defined as the abilities to self-renew and to generate differentiated progeny, which give rise to the histologically heterogeneous appearance of tumors. In addition to primary tumors, immortalized cancer cell lines are widely used as an experimental model for CSC research due to the ease of maintaining these cells and their limitless proliferation potential in culture.(4) Although immortalized cancer cell lines have been a valuable resource for understanding the molecular biology of solid tumors, it is unclear whether information gleaned from the study of CSCs in primary tumors can be applied to the biology of cancer cell lines. In fact, some cell lines lack a small population of cells that display accepted CSC markers, such as a side population that is not labeled with the fluorescent dye Hoechst 33342, and the expression of CD133.(5,6) However, even in cell lines that lack traditional CSC markers, CSCs are believed to be present, because these lines have been maintained over decades and retain their tumorigenic potential. An important unanswered question is whether CSC markers identified in primary cancer specimens can also be applied to immortalized cancer cell lines.

Sphere initiation from single cells in suspension is a hallmark of CSCs, as well as normal neural and mammary stem cells.(7–12) Similar to the behavior of tissue stem cells in floating culture in the presence of epidermal growth factor (EGF) and basic-fibroblast growth factor (b-FGF), primary cancer stem/progenitor cells expand within floating-spheres.(10,12,13) This method has been used successfully to enrich for CSCs, which are scarce and difficult to maintain in vitro.(14,15) In theory, applying the behavior of CSCs from primary tumors directly to cancer cell lines, sphere-formation assays using cancer cell lines could provide an attractive opportunity to screen drugs that target CSCs. To date, however, reports of the significance of sphere-formation in cancer cell lines and its clinical relevance are scarce.

A consensus regarding CSC hierarchy in cancer cell lines is needed. In particular, several outstanding questions remain to be addressed, including the following: (i) are cell-surface markers identified on primary tumors appropriate for the analysis of cancer cell lines; (ii) is there a CSC hierarchy within some or all cancer cell lines; and (iii) what is the biological and clinical significance of sphere-forming activity in cancer cell lines?

Here we present the results of our analysis of the CSC hierarchy of the HCT116 colon cancer cell line. HCT116 cells did not self-renew from single cells under floating culture conditions, but were able to propagate once aggregated-spheres were allowed to form. Analysis of the expression of epithelial specific antigen (ESA) and CD44, which are markers of colon CSCs, indicated that sphere-formation induces a phenotypic hierarchy within HCT116 cells, which was represented by two main subpopulations, ESAlowCD44−/low and ESAhighCD44high. The phenotypic hierarchy was re-established by either subpopulation once the cells were cultured in aggregative and cell-dense conditions in floating culture. These results suggest that within the HCT116 cell population, all cells possess the plasticity required to initiate tumors, and that the maintenance of this cancer cell line conforms to the stochastic model, not the CSC model.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Cell culture.  The colon cancer cell lines WiDr, HT29, DLD-1, and HCT116 were cultured in DMEM/F12 media supplemented with 10% FCS, 100 units/mL penicillin G, and 100 μg/mL streptomycin (Nacalai tesque, Kyoto, Japan). For floating culture, HCT116 cells were cultured in serum-free DMEM/F12 media containing 5 μg/mL of bovine insulin, 0.4% BSA, 10 ng/mL of b-FGF, and 20 ng/mL of EGF in 100 mm ultra-low attachment dishes (Corning Japan, Tokyo, Japan) at a density of 2.0 × 104 cells/mL at 37°C in a humidified 5% CO2/95% air atmosphere.

Flow cytometry.  The expression of ESA, CD44, and CD133 was analyzed in cells derived from monolayer cultures, or 4-day-old primary floating-spheres following dissociation by NeuroCult (Stemcell Technologies, Vancouver, BC, Canada). Cells (5−10 × 105) were pelleted by centrifugation at 500g for 3 min at 4°C and then resuspended in 100 μL of PBS. Cell suspensions were added to 2.5 μL of FITC-conjugated mouse monoclonal antihuman ESA Ab (BD Pharmingen, San Jose, CA, USA), 5.0 μL of Allophycocyanin (APC)-conjugated mouse monoclonal antihuman CD44 Ab (BD Pharmingen), and 10 μL of Phycoerythrin (PE)-conjugated mouse monoclonal antihuman CD133 Ab (BD Pharmingen). Suspensions were incubated for 30 min at 4°C. At least three independent experiments were performed. Cells were washed with a solution of PBS containing 3% FCS. FACS analysis and sorting were performed using a FACSVantage SE cell sorter (Becton Dickinson, Franklin Lakes, NJ, USA). To analyze cell-cycle status, cells were first stained with Hoechst 33342 at 37°C for 45 min, after which 1 μg/mL of Pyronin Y was added, and then the cells were incubated at 37°C for 45 min. Dead cells were depleted by propidium iodide prior to FACS analysis.

Floating-sphere formation assay.  The sphere-forming activity of HCT116 cells derived from monolayer culture or floating culture was assessed by plating cells into a 24-well dish at a density of 200−3000 cells/well. Floating-spheres were counted 3 days after plating.

Colony-formation in soft agar assay.  To evaluate anchorage-independent cell survival and proliferation, we performed a colony-formation assay. A bottom layer of 0.5 × DMEM/F12 media + 0.5% agar was poured into 1.5 mL/3.5 cm Petri dishes. After the agar solidified, a top layer of 0.5 × DMEM/F12 media + 0.35% agar mixed with a specified number of cells (5000 cells/dish) was added to the dish. The dishes were incubated at 37°C in a humidified incubator for 14 days. Colonies were stained with 0.5 mL of 0.005% crystal violet for at least 1 h, and then counted using a dissecting microscope.

Immunostaining of sphere cells.  Spheres were harvested after 4 days in floating culture, fixed with 4% paraformaldehyde or 95% ethanol for 10 min at room temperature, and then embedded in optimal cutting temperature (OCT) compound (Miles Laboratories, Elkhart, IN, USA). Serial sections (4 mm) were generated and then stained with one of the following Abs, as indicated: mouse monoclonal anti-CD326 (ESA) (1:10; Abcam, Cambridge, UK), rat monoclonal anti-CD44 (1:100; Santa Cruz Biotechnology, Santa Cruz, CA, USA), mouse monoclonal anti-E-cadherin (1:100; BD Biosciences, Franklin Lakes, NJ, USA), mouse monoclonal anti-pan-p38 (1:100; Santa Cruz Biotechnology), rabbit monoclonal anti-phospho-p38 (1:100; Cell Signaling, Trask Lane, Danvers, USA), rabbit polyclonal anti-APG8b (1:100; ABGENT, San Diego, CA, USA), or mouse monoclonal anti-BrdU (1:50; Dako, Glostrup, Denmark). Primary antibodies were detected using Alexa488-conjugated goat antimouse IgG, Texas Red-conjugated goat antirat IgG (1:100), or Texas Red-conjugated goat antirabbit IgG (1:100) (Invitrogen, Carlsbad, CA, USA). The cells were counterstained with DAPI to identify nuclei, and then viewed under a fluorescence microscope (BIOREVO BZ-9000; Keyence, Osaka, Japan). For immunohistochemical staining, after incubation with primary Abs, sections were incubated at 37°C for 30 min with HRP–labeled secondary goat antimouse or antirabbit IgG. Sections were stained with diaminobenzidine, counterstained with hematoxylin to visualize the cell nuclei, and then examined by microscopy. For the BrdU incorporation assay, spheres were incubated for 4 h with BrdU (2.5 mm) and then sections were analyzed by immunostaining using anti-BrdU Ab.

Transplantation into NOD/SCID mice.  Non-obese diabetic (NOD)/SCID mice were purchased from Oriental Yeast (Tokyo, Japan). The indicated number of floating-sphere-derived ESAlowCD44low/− and ESAhighCD44high cells were suspended in a solution of DMEM/F12 and Matrigel (1:1; Becton Dickinson), and then transplanted into 6–10-week-old NOD/SCID male mice under anesthesia. ESAlowCD44low/− and ESAhighCD44high cells were implanted into the subcutaneous space of the left and right side of the back, respectively. Tumor formation was observed weekly for 12 weeks. Transplantation assays were performed in accordance with the institutional guidelines of the Keio University Animal Committee.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

CD133 is not an appropriate CSC marker for HCT116 cells.  We examined several cell lines for the presence of subpopulations of cells that expressed CD133, which is a generally accepted CSC marker in primary colon cancer.(16,17) Consistent with a previous report,(18) >70% of WiDr and HT29 colon cancer cells were positive for CD133 (CD133+) (Fig. 1a). In contrast, approximately 1% of DLD1 cells were CD133+ (Fig. 1a), and CD133+ cells were undetectable in the HCT116 cell line. To rule out the possibility that the number of CD133+ HCT116 cells was below the limit of detection of FACS analysis, we carried out a floating-sphere formation assay to enrich for CSCs (Fig. 1b). CD133+ HCT116 cells were undetectable in sphere-derived cells (Fig. 1c), which indicated that CD133 is not an appropriate CSC marker for HCT116 cells. These results suggested that there are other CSC markers that might be more appropriate for the purification of CSCs in the HCT116 line, as in primary colon cancer.(19) To this end, we focused the remaining experiments on the HCT116 cell line.

image

Figure 1.  CD133 is not an appropriate cancer stem cell (CSC) marker for HCT116 cells. (a) Expression of CD133 in the human colon cancer cell lines WiDr, HT29, DLD-1, and HCT116. Dashed line, isotype control Ab. PE, phycoerythrin. (b) HCT116 cell spheres in floating cultures. (c) FACS analysis of CD133 expression in sphere-derived HCT116 cells. SCC-A, side scatter-area.

Download figure to PowerPoint

Sphere-formation in floating culture induces a biological/proliferative hierarchy, but not a CSC hierarchy, within HCT116 cells.  To determine whether there is a CSC hierarchy in HCT116 cells, we examined the expression of the alternative CSC markers CD166, ESA, and CD44. We could not demonstrate significant heterogeneity in the expression of CD166 in HCT116-adherent cells. HCT116 cells were all positive for CD166 (data not shown). Thus, we focused on ESA and CD44 expression as other candidate markers for CSCs. Tumor-initiating cells are reportedly enriched in the fraction of primary colon cancer cells that express high levels of ESA and CD44 (ESAhighCD44+).(19) We examined CSC marker expression in floating cultures as well as adherent conditions as a way to enrich for the ESAhighCD44+ CSC subpopulation in HCT116 cells, since floating culture has been used successfully to enrich for the CD133+ subpopulation in primary colon cancers.(12,13)

Adherent cells fractionated into two subpopulations, ESAhighCD44−/low and ESAhighCD44high (Fig. 2a, left panel), whereas floating-sphere derived cells fractionated into ESAlowCD44−/low and ESAhighCD44high subpopulations, which indicated that ESA expression is dependent on culture conditions (Fig. 2a, right panel). To determine whether the differences in ESA and CD44 expression between adherent and floating cells were associated with altered tumorigenic profiles, we compared the sphere-forming activity of the two main fractions derived from adherent and floating cultures, as sphere-formation has been shown to correlate with tumor-initiating potential in vivo.(20) There was no difference in the sphere-forming activity between the ESAhighCD44−/low and ESAhighCD44high fractions of adherent cells (data not shown). In contrast, the ESAhighCD44high fraction of floating-sphere-derived cells had a higher sphere-forming activity than the ESAlowCD44−/low fraction (Fig. 2b, left panel). We also analyzed the fractions using a colony-formation in soft agar assay, which provides a measure of the anchorage-independent proliferative activity of transformed cells in vitro. The ESAhighCD44high fraction of floating-sphere derived cells had a higher colony-formation activity than the ESAlowCD44−/low fraction (Fig. 2b, right panel), whereas there was no difference in colony-formation activity between the two major fractions of adherent cells (data not shown). These results indicated that there is no proliferative hierarchy in adherent HCT116 cells, whereas under floating culture conditions, HCT116 cells exhibit a hierarchy of sphere-forming and colony-forming activity. We hereafter refer to this hierarchy within sphere-derived HCT116 cells as ‘biological or proliferative hierarchy’.

image

Figure 2.  Floating culture sphere-formation induces a biological/proliferative, but not cancer stem cell (CSC) hierarchy, in HCT116 cells. (a) Representative FACS dot plots showing the expression of epithelial specific antigen (ESA) and CD44 in adherent and sphere-derived HCT116 cells. APC, allophycocyanin. (b) Analysis of sphere-formation (left) and colony-formation in agar (right) by FACS-sorted subpopulations of sphere-derived HCT116 cells. (c) Representative subcutaneous tumors were derived from 5.0 × 103 ESAlowCD44−/low (left, arrow) and ESAhighCD44high (right, arrow) HCT116 sphere cells. Corresponding sections were stained with H&E (upper) and subjected to immunohistochemical analysis with anti-CD44 (middle) and anti-ESA (lower) Abs (magnification, ×40). (d) Volume of tumors derived from 5.0 × 103 cells from each subpopulation, as described for (c), 8 weeks after inoculation. Data represents the means ± SEM of four mice inoculated as one series of inoculations.

Download figure to PowerPoint

To determine whether the biological/proliferative hierarchy of HCT116 cells was related to CSC hierarchy, we carried out a xenotransplantation assay. CSC hierarchy is defined as the hierarchy of tumor incidence among definite subpopulations of bulk tumor cells, and also as the potential to reconstitute the pathological features of primary tumors. It is not simply a hierarchy of differences in proliferative activity.(3,21,22) In the xenograft transplant model, a high incidence of tumors derived from a limited number of cancer cells indicates that the fraction is a CSC-enriched subpopulation. We transplanted floating-sphere-derived fractions of HCT116 cells into NOD/SCID mice. At a transplant cell density of 5.0 × 103 cells per inoculation, tumors derived from ESAhighCD44high cells grew much faster than those from ESAlowCD44−/low cells (Fig. 2c,d). However, tumor incidence between the two subpopulations was similar (Table 1). At a cell density of 200 cells per inoculation into the flanks of NOD/SCID mice, the tumor incidence of the ESAlowCD44−/low fraction was three per three injections, and that of the ESAhighCD44high fraction was two per three injections. An inoculum of 200 cells is generally regarded as the limiting density for tumor initiation for primary colon cancer cells.(19) Thus, there was no evidence of a CSC hierarchy within HCT116 sphere-derived cells. Of note, the reconstituted tumors derived from either subpopulations exhibited the same histological phenotypes as undifferentiated carcinomas and displayed similar expression patterns of CD44 and ESA (Fig. 2c).

Table 1.   Incidence of tumors derived from distinct subpopulations in HCT116 sphere cells. Cell doses and tumor formation†
Phenotypic subpopulations4.0 × 1042.5 × 1041.0 × 104500030001000500200
  1. †Cell dose, number of cells per injection; tumor formation, number of tumors formed/number of injections; tumor take was considered unsuccessful when no tumor mass was palpable after 2 months’ follow-up. ‡For every seven injections, four injections were performed as one series of inoculations. Four tumors derived from each inoculation series were analyzed for their volumes (Fig. 2d).

ESAloCD44−/lo1/11/13/37/7‡2/33/33/33/3
ESAhiCD44hi1/11/12/36/7‡2/33/33/32/3

Overall, both in vitro and in vivo results indicated that floating culture conditions induced a biological/proliferative hierarchy, but not a CSC hierarchy, within the HCT116 cell population.

Phenotypic hierarchy of HCT116 spheres is regulated by the microenvironment, but not by extrinsic factors.  We performed a limiting dilution analysis in conjunction with the floating-sphere assay to determine whether the biological hierarchy within sphere-derived HCT116 cells was due to heterogeneity in the self-renewing potential of individual cells. At a density of <200 cells/mL, no floating-spheres were generated, even by the ESAhighCD44high fraction (Fig. 3a). These results indicated that HCT116 cells are incapable of proliferating or self-renewing from a single cell in floating culture. Thus, sphere initiation is not due to the self-renewing potential of individual cells; rather, it is dependent on cellular aggregation. To determine whether the biological/proliferative hierarchy of HCT116 sphere cells was regulated by extrinsic factors, we examined the effect of EGF and b-FGF, which have been shown to promote the self-renewal of primary colon CSCs, on floating-sphere formation. The phenotypic hierarchy, represented by ESAlowCD44−/low and ESAhighCD44high, was repopulated in serum-containing medium in the presence and absence of EGF and b- FGF (Fig. 3b), which indicated that the hierarchy is not regulated by extrinsic factors, but rather, is related to the structural features of floating-spheres.

image

Figure 3.  HCT116 sphere-formation is due to cellular aggregation and exhibits a stable phenotype, independent of extrinsic growth factors. (a) Sphere-formation assay of fractionated sphere-derived HCT116 cells plated at the indicated cell densities. Data represents the means ± SEM of at least three wells for each density. ESAhighCD44high is underlined to distinguish it clearly from ESAlowCD44−/low. (b) FACS analysis of sphere-derived HCT116 cells cultured under floating culture conditions in the presence of epidermal growth factor (EGF) and basic-fibroblast growth factor (b-FGF) (top right), EGF, bFGF, and FCS (bottom left), or FCS (bottom right). The phenotype of adherent HCT116 cells cultured in the presence of 10% FCS was analyzed as a control (top left).

Download figure to PowerPoint

The biological/proliferative hierarchy within HCT116 sphere-derived cells is closely related to structural cues.  To analyze the role of structural cues in the phenotypic hierarchy of sphere-derived HCT116 cells, we analyzed the spheres by immunofluorescence and immunohistochemical staining. ESA and CD44 were highly expressed in the outer layers of the spheres, and the intensity of staining decreased toward the core (Fig. 4a). Sites of cell–cell contact, which can be detected as regions of colocalized E-cadherin and β-catenin, were absent and cell density was lower in the core of the sphere as compared to the outer layer (Fig. 4a). In contrast, markers of cellular stress and autophagy (phosphorylated-p38 and APG8b, respectively) were strongly expressed within the inner layer of the sphere (Fig. 4b). These results suggested that the phenotypic hierarchy of sphere-derived HCT116 cells is due in part to a gradient of cellular stress or nutrient status within each sphere, and most likely does not reflect different levels of differentiation or intrinsic cellular properties.

image

Figure 4.  Structural and biological characterization of HCT116 spheres. (a) Expressions of epithelial specific antigen (ESA) and CD44 (top), and E-cadherin and β-catenin (bottom) in HCT116 spheres were analyzed by fluorescence microscopy. Nuclei were counterstained with DAPI (blue staining). (b) Immunohistochemical analysis of frozen sections of HCT116 spheres using anti-phospho-p38 (top left), anti-pan-p38 (top right), anti-APG8b (bottom left), and anti-BrdU (bottom right) Abs. (c) Cell-cycle analysis of HCT116 sphere-derived cells. Representative staining pattern for ESA and CD44 in sphere-derived HCT116 cells (top left) is shown. Hoechst 33342 and Pyronin Y staining pattern of intact sphere-derived HCT116 cells (top right), and fractionated ESAlowCD44−/low (bottom left, arrow) and ESAhighCD44high (bottom right, arrow) cells are also shown.

Download figure to PowerPoint

When we analyzed the cell-cycle status of ESAlowCD44−/low and ESAhighCD44high fractions, ESAhighCD44high cells were mainly in G1 (91.2%), with a small percentage in S/G2M (4.58%), and no cells in G0 (Fig. 4c). By comparison, approximately 20% of ESAlowCD44−/low cells resided in G0, 66% were in G1, and 0.69% were in S/G2M. These findings were consistent with the results of the BrdU incorporation assay of HCT116 spheres, which indicated that cells located in the outer layer, but not the inner layer, were actively proliferating (Fig. 4b). These results suggested that there are cells within the core of HCT116 spheres that are dormant as a result of cellular stress or starvation. Furthermore, these results are inconsistent with the characteristics of differentiated progeny and quiescent CSCs. Thus, the biological/proliferative hierarchy within sphere-derived HCT116 cells appears to derive from the effects of cell localization within the sphere, rather than autonomous CSC properties.

Phenotypic conversion of suspended cells following aggregation.  Given that in floating culture, a single HCT116 cell was incapable of proliferating (Fig. 3a), we investigated whether cellular aggregation was a key event in establishing the biological/proliferative and phenotypic hierarchy of ESA−/low CD44low and ESAhighCD44high HCT116 cells. Sphere-derived cells were allowed to form cellular aggregates in floating culture, and then ESA and CD44 expression was examined to determine whether the phenotypic hierarchy was re-established in cellular aggregates derived from either subpopulation. As shown in Figure 5, the phenotypic hierarchy of ‘ESA−/low CD44low and ESAhighCD44high’ was re-established by both subpopulations following aggregation.

image

Figure 5.  Reconstitution of the phenotype of sphere-derived HCT116 cells following aggregation of ESAlowCD44−/low and ESAhighCD44high subpopulations. Representative FACS dot plots of (a) intact sphere-derived HCT116 cells at the time of sorting and fractionated (b) ESAlowCD44−/low and (c) ESAhighCD44high subpopulations after culturing at an aggregative density of 5000 cells/mL. APC, allophycocyanin.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

The notion that tumors are composed of a heterogeneous population of cells has been an empirically established concept and has been accepted by the medical field for some time.(3) However, starting in the 1960s, evidence began emerging that challenged the traditional notion of cancer cell biology, and ultimately created a new field of cancer biology, the field of CSC research.(21) In mice with malignant myeloma, only a subset of intact malignant cells form colonies in vitro, at a rate of 1 in 10 000 to 1 in 100 malignant cells.(23) Furthermore, when mouse lymphoma cells are transplanted in vivo, approximately 1% of the cells form spleen colonies.(24) Based on these and other similar findings, two conflicting models have emerged to explain heterogeneity in the colony-forming potential of malignant cells: the stochastic model and the CSC model.(21) In primary tumors, including hematopoietic malignancies, the CSC model has been well established using cell isolation techniques based on putative CSC markers and xenotransplantation.(1,2) For primary tumors with a clear hierarchy of tumor incidence among different tumor cell fractions,(16,17) the CSC hypothesis is well supported. However, in cancer cell lines, differences in tumor incidence among subpopulations that are isolated based on the expression of CSC markers are not as clearly defined.(18) For example, most C6 rat glioma cells can proliferate and exhibit tumorigenic properties in clonal and cell-based assays.(25) Thus, it remains to be determined whether cancer cell lines adhere to a stochastic or CSC model of tumorigenesis.

In the current study, we investigated whether traditional CSC markers, based on studies of primary tumors, are appropriate for the analysis of cancer cell lines. In the HCT116 colon cancer cell line, we were unable to identify a CD133+ subpopulation. This result conflicts with previous reports,(6,26) in which the HCT116 line was shown to have a CD133+ subpopulation. One explanation for this inconsistency could be phenotypic differences between the HCT116 clones used in the two studies, which might have occurred over long periods of cell culture. Alternatively, this discrepancy could be due to differences in the definition of what constitutes the threshold of CD133 positivity. We set the threshold strictly at the upper end of the fluorescent distribution of the cells stained with PE-conjugated isotype IgG. In fact, previous reports showed that CD133+ subpopulations among HCT116 cells are rare, representing 0.3 ± 0.1% and around 3% or less.(6,26) In addition, we could not detect CD133+ HCT116 cells even among sphere-derived cells, suggesting that a CD133+ subpopulation is not required for the maintenance of the cell line we used.

Given that CD133 CSCs have been reported in glioblastoma,(27) we cannot rule out the possibility that HCT116 cells are maintained by CSCs that express other CSC markers, such as ESA and CD44. In the prevailing CSC model, CSCs are enriched in floating-sphere cultures in serum-free medium supplemented with EGF and b-FGF.(10–15) Interestingly, when we examined the expression patterns of ESA and CD44 as candidate alternative CSC markers(19) in adherent and floating cultures of HCT116 cells, we observed the differences in ESA and CD44 expression between adherent and sphere-derived cells. Further analysis revealed that these phenotypic differences in CSC marker expression in sphere-derived cells reflected a biological/proliferative hierarchy in vitro, namely in sphere-forming and colony-forming activity. To determine whether the biological/proliferative hierarchy of sphere-derived cells corresponded to a CSC hierarchy, we performed a limiting dilution analysis in conjunction with the floating-sphere assay, and a xenotransplantation assay. In floating culture, at a cell density of <200 cells/mL, neither fraction of sphere-derived HCT116 cells developed spheres. In the xenotransplantation assay, although tumor growth was significantly different between the two fractions, tumor incidence was the same. Tumor initiating activity, which dictates tumor incidence, does not correspond to the proliferating activity of tumor cells. In addition, the tumors derived from each subpopulation had the histological features of undifferentiated monoclonal-like tumors. These findings suggest that the biological/proliferative hierarchy in vitro of sphere-derived HCT116 cells is not due to autonomous cell properties, that is CSC properties, but rather to the proliferative capacity of the cells, and that cellular aggregation is essential for the proliferation of HCT116 cells in vitro.

Immunohistochemical analysis of HCT116 spheres revealed that cell–cell contacts were dissolved, and that the expression of ESA and CD44 was reduced in the core of the spheres as compared to the outer layer. Furthermore, the levels of the autophagic marker APG8b and phosphorylation of the stress-activated protein kinase p38 were higher in cells in the core region. Cell-cycle analysis of spheres derived from ESA−/low CD44low fraction indicated the presence of dormant cells in G0 stage. We speculate that this phenotypic gradient is due to a gradient of cellular stresses, such as starvation and hypoxia, and reflects the location of cells within the sphere. It is important to note here that the dormant cells in the core of the sphere had not undergone senescence following differentiation, because ESA−/lowCD44low cells were able to reconstitute the phenotypic hierarchy of HCT116 cells (ESA−/lowCD44low and ESAhighCD44high) once they formed aggregates. Thus, individual HCT116 cells have the potential to proliferate once they aggregate in floating culture, which indicates that the maintenance of the HCT116 cell line conforms to a stochastic model, not a CSC model. In another colon cancer cell line, DLD-1, we also observed a change in the hierarchy of ESA and CD44 phenotypic expression. This change depended on whether adherent or floating culture conditions were used (data not shown). Therefore, it is possible to apply this notion to some other cancer cell lines.

The results of the current study are inconsistent with a previous report by Locke et al.(28) However, it is important to note that the work of Locke et al., as well as that of Zheng et al.,(25) did not include FACS analysis for the identification of CSCs. Rather, definitive hierarchies of colony-forming activity within in some cell lined were identified by re-plating morphologically distinct colonies, and evaluating colony-forming potential based on the morphology of the parental colonies. Further analysis to determine whether this hierarchy in colony-formation activity is related to CSC hierarchy is warranted.

Previously, it was reported that the expression of the CSC marker CD133 oscillates with cell cycle progression in human embryonic stem cell, colon cancer, and melanoma cell lines.(29) Thus, the idea of cancer cells’‘plasticity’, and not CSC properties, has some precedence. In cancer cell lines, the expression of CSC markers at any given moment is a snapshot of the evolving state of the cells. Therefore, we here rephrase this snapshot of the hierarchy of the cells as ‘plastic hierarchy’, which incorporates the concept of cancer cell lines as dynamic and stochastic.

In conclusion, the novel results of the current study elucidate a mechanism of cancer cell line maintenance that conforms more to the stochastic model than to the CSC model. Our results have two main implications: (i) the use of certain cancer cell lines for CSC research may not be appropriate; and (ii) some primary cancer cells might also exhibit plasticity in terms of self-maintenance. Thus, certain stages or kinds of malignant tumors need to be analyzed in terms of the stochastic model, as recently suggested for human melanomas.(30) As the distinction between the two models ultimately impacts strategies for the development of anticancer drugs, further studies to resolve the controversy are warranted.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

We thank the Department of Biomedical Research and Development, Link Genomics, Inc., for the gift of WiDr, HT29, and DLD1 colon cancer cell lines, and members of the Division of Gene Regulation, Institute for Advanced Medical Research, School of Medicine, Keio University, for helpful discussions. This study was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (H.S.).

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References
  • 1
    Bonnet D, Dick JE. Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med 1997; 3: 7307.
  • 2
    Visvader JE, Lindeman GJ. Cancer stem cells in solid tumours: accumulating evidence and unresolved questions. Nat Rev Cancer 2008; 8: 75568.
  • 3
    Dick JE. Stem cell concepts renew cancer research. Blood 2008; 112: 4793807.
  • 4
    Kondo T. Stem cell-like cancer cells in cancer cell lines. Cancer Biomark 2007; 3: 24550.
  • 5
    Kondo T, Setoguchi T, Taga T. Persistence of a small subpopulation of cancer stem-like cells in the C6 glioma cell line. Proc Natl Acad Sci U S A 2004; 101: 7816.
  • 6
    Yi JM, Tsai HC, Glockner SC et al. Abnormal DNA methylation of CD133 in colorectal and glioblastoma tumors. Cancer Res 2008; 68: 8094103.
  • 7
    Reynolds BA, Weiss S. Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 1992; 255: 170710.
  • 8
    Reynolds BA, Weiss S. Clonal and population analyses demonstrate that an EGF-responsive mammalian embryonic CNS precursor is a stem cell. Dev Biol 1996; 175: 113.
  • 9
    Dontu G, Abdallah WM, Foley JM et al. In vitro propagation and transcriptional profiling of human mammary stem/progenitor cells. Genes Dev 2003; 17: 125370.
  • 10
    Ponti D, Costa A, Zaffaroni N et al. Isolation and in vitro propagation of tumorigenic breast cancer cells with stem/progenitor cell properties. Cancer Res 2005; 65: 550611.
  • 11
    Bar EE, Chaudhry A, Lin A et al. Cyclopamine-mediated hedgehog pathway inhibition depletes stem-like cancer cells in glioblastoma. Stem Cells 2007; 25: 252433.
  • 12
    Vermeulen L, Todaro M, De Sousa Mello F et al. Single-cell cloning of colon cancer stem cells reveals a multi-lineage differentiation capacity. Proc Natl Acad Sci U S A 2008; 105: 1342732.
  • 13
    Todaro M, Alea MP, Di Stefano AB et al. Colon cancer stem cells dictate tumor growth and resist cell death by production of interleukin-4. Cell Stem Cell 2007; 1: 389402.
  • 14
    Ricci-Vitiani L, Pagliuca A, Palio E, Zeuner A, De Maria R. Colon cancer stem cells. Gut 2008; 57: 53848.
  • 15
    Cammareri P, Lombardo Y, Francipane MG, Bonventre S, Todaro M, Stassi G. Isolation and culture of colon cancer stem cells. Methods Cell Biol 2008; 86: 31124.
  • 16
    O’Brien CA, Pollett A, Gallinger S, Dick JE. A human colon cancer cell capable of initiating tumour growth in immunodeficient mice. Nature 2007; 445: 10610.
  • 17
    Ricci-Vitiani L, Lombardi DG, Pilozzi E et al. Identification and expansion of human colon-cancer-initiating cells. Nature 2007; 445: 1115.
  • 18
    Ieta K, Tanaka F, Haraguchi N et al. Biological and genetic characteristics of tumor-initiating cells in colon cancer. Ann Surg Oncol 2008; 15: 63848.
  • 19
    Dalerba P, Dylla SJ, Park IK et al. Phenotypic characterization of human colorectal cancer stem cells. Proc Natl Acad Sci U S A 2007; 104: 1015863.
  • 20
    Patrawala L, Calhoun T, Schneider-Broussard R et al. Highly purified CD44+ prostate cancer cells from xenograft human tumors are enriched in tumorigenic and metastatic progenitor cells. Oncogene 2006; 25: 1696708.
  • 21
    Reya T, Morrison SJ, Clarke MF, Weissman IL. Stem cells, cancer, and cancer stem cells. Nature 2001; 414: 10511.
  • 22
    Dick JE. Looking ahead in cancer stem cell research. Nat Biotechnol 2009; 27: 446.
  • 23
    Park CH, Bergsagel DE, McCulloch EA. Mouse myeloma tumor stem cells: a primary cell culture assay. J Natl Cancer Inst 1971; 46: 41122.
  • 24
    Bruce WR, Van Der Gaag H. A quantitative assay for the number of murine lymphoma cells capable of proliferation in vivo. Nature 1963; 199: 7980.
  • 25
    Zheng X, Shen G, Yang X, Liu W. Most C6 cells are cancer stem cells: evidence from clonal and population analyses. Cancer Res 2007; 67: 36917.
  • 26
    Botchkina IL, Rowehl RA, Rivadeneira DE et al. Phenotypic subpopulations of metastatic colon cancer stem cells: genomic analysis. Cancer Genomics Proteomics 2009; 6: 1929.
  • 27
    Beier D, Hau P, Proescholdt M et al. CD133(+) and CD133(−) glioblastoma-derived cancer stem cells show differential growth characteristics and molecular profiles. Cancer Res 2007; 67: 40105.
  • 28
    Locke M, Heywood M, Fawell S, Mackenzie IC. Retention of intrinsic stem cell hierarchies in carcinoma-derived cell lines. Cancer Res 2005; 65: 894450.
  • 29
    Jaksch M, Munera J, Bajpai R, Terskikh A, Oshima RG. Cell cycle-dependent variation of a CD133 epitope in human embryonic stem cell, colon cancer, and melanoma cell lines. Cancer Res 2008; 68: 78826.
  • 30
    Quintana E, Shackleton M, Sabel MS, Fullen DR, Johnson TM, Morrison SJ. Efficient tumour formation by single human melanoma cells. Nature 2008; 456: 5938.